BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//talks.bham.ac.uk//v3//EN
BEGIN:VEVENT
CATEGORIES:Data Science and Computational Statistics Seminar
SUMMARY:Machine learning for coarse-graining molecular sys
tems - Gabriel Stoltz (Ecole des Ponts and Inria P
aris)
DTSTART:20211027T120000Z
DTEND:20211027T130000Z
UID:TALK4678AT
URL:/talk/index/4678
DESCRIPTION:A coarse-grained description of atomistic systems
in molecular dynamics is provided by reaction coor
dinates. These nonlinear functions of the atomic p
ositions are a basic ingredient to compute more ef
ficiently average properties of the system of inte
rest\, such as free energy profiles. However\, rea
ction coordinates are often based on an intuitive
understanding of the system\, and one would like t
o complement this intuition or even replace it wit
h automated tools. One appealing tool is autoencod
ers\, for which the bottleneck layer provides a lo
w dimensional representation of high dimensional a
tomistic systems. I will discuss some mathematical
foundations of this method\, and present illustra
tive applications including alanine dipeptide and
chignolin. Some on-going extensions to more demand
ing systems\, namely HSP90\, will also be hinted a
t. Joint work with Zineb Belkacemi (Sanofi and Eco
le des Ponts)\, Tony LeliÃ¨vre (Ecole des Ponts and
Inria Paris)\, Gkeka Paraskevi (Sanofi).
LOCATION:Seminar Room 119\, Arts Building (R16) or https:/
/bham-ac-uk.zoom.us/j/95645638485?pwd=eDh3dlR3QjRl
VWRHdExvaDBWdEo4UT09
CONTACT:Hong Duong
END:VEVENT
END:VCALENDAR